s: Improving cotton yield and fiber quality is a major breeding goal for Chinese upland cotton. Discovery molecular markers associated with the target traits can overcome the blindness of conventional breeding and improve the accuracy of molecular marker assisted selection breeding. In this study, a diversity panel consisting of 118 upland cotton accessions was grown in four different environments, and the yield-related traits including lint percentage (LP), boll weight (BW), bolls per plant (BN) and seed index (SI) were measured, and 214 polymorphic SSR markers of covering the whole genome were performed in association mapping. The results were showed: 118 upland cotton accessions were rich in phenotypic variation, the mean coefficients of variance (CV) ranged from 6.1~19.1%, and were more stable in different environments; Genotyping was performed with genome-wide Simple Sequence Repeats (SSRs), a total of 201markers were polymorphic and generated 460 allele loci, the average genetic diversity index and polymorphism information content (PIC) were 0.5151 and 0.4587, respectively. The results indicated that the alleles and the genetic diversity were larger in upland cotton accessions; 118 cotton varieties were divided into 4 subgroups by analysis of population structure, there has no corresponding relation between each kind of group of accessions and the geographical regions. Association analysis results were showed that 39 marker loci were detected at least two of the four environments (-lg(p) >1.3, P<0.05 ), Of 39 associated markers, four marker loci were associated with more than two traits simultaneously, seven marker loci were coincident with previous studies, and 32 were identified as new association markers. The results could provide a theoretical basis for molecular assisted selection of upland cotton about yield traits.